Knee osteoarthritis is a major cause of chronic pain, mobility restriction, and loss of independence in older adults. When advanced structural disease is accompanied by persistent symptoms despite appropriate non-operative treatment, total knee arthroplasty (TKA) is an established intervention for relieving pain and restoring function. Contemporary evidence confirms that knee replacement is among the most frequently performed and clinically effective musculoskeletal procedures, with most patients achieving substantial improvements in pain, activities of daily living, and health-related quality of life [1]. Nevertheless, recovery after TKA is not uniform. A clinically important subgroup continues to experience pain, functional limitation, or dissatisfaction despite technically successful surgery, and systematic reviews have reported unfavorable pain outcomes in a meaningful proportion of patients after knee replacement [2].

Functional recovery following TKA is influenced by the interaction of preoperative patient characteristics, baseline physical performance, comorbidity burden, surgical stress, postoperative pain, and rehabilitation. Earlier prospective work demonstrated that poor baseline function, severe symptoms, and a greater number of comorbid conditions predict inferior postoperative outcomes [3,4]. During the acute recovery phase, older age, comorbidity, and greater pain have also been associated with poorer physical performance [5]. These findings emphasize that implant success alone does not determine recovery; patient-level factors and early postoperative events contribute substantially to the speed and extent of functional restoration.

Identifying patients at risk of delayed recovery before or immediately after surgery is clinically valuable. Prediction can support realistic counselling, targeted prehabilitation, optimization of metabolic and hematologic status, intensified pain control, and individualized rehabilitation. Clinical prediction research has linked higher body mass index, poorer preoperative physical status, and worse patient-reported function with less favorable outcomes after knee replacement [6]. Performance-based measures provide complementary information. Standardized assessments incorporating the Timed Up and Go (TUG) test and walking capacity have shown that slower preoperative mobility is associated with poorer postoperative function [7]. In parallel, greater acute postoperative pain has been associated with lower subsequent functional scores and satisfaction [8], while systematic evidence indicates that perioperative pain and patient-level factors contribute to persistent postoperative symptoms [9].

Despite these observations, the relative contribution of readily measurable clinical predictors can differ according to population, perioperative pathway, rehabilitation practices, and the time point used to define recovery. Short-term identification of delayed functional recovery is particularly relevant because the first three postoperative months represent an active rehabilitation period during which modifiable barriers can be addressed. Therefore, this prospective observational study was undertaken to determine the proportion of patients with delayed functional recovery 12 weeks after primary TKA for advanced osteoarthritis and to evaluate demographic, metabolic, preoperative functional, and early perioperative factors associated with delayed recovery. The primary objective was to identify independent predictors of delayed functional recovery; secondary objectives were to compare serial WOMAC function, TUG performance, knee flexion, ambulation status, and perioperative characteristics between patients with timely and delayed recovery.

Methodology Study design and setting. This prospective observational study was conducted in the Department of Orthopaedics at TRR Institute of Medical Sciences, Patancheru, Sangareddy, Telangana, India, from January 2026 to June 2026. Consecutive patients undergoing primary unilateral total knee arthroplasty (TKA) for advanced osteoarthritis were screened, and recruitment was completed early enough to permit 12-week follow-up within the study period. Study participants. Adults aged 50 years or older with advanced primary knee osteoarthritis, persistent pain and functional limitation despite conservative treatment, and planned primary TKA were eligible. Exclusion criteria were revision or simultaneous bilateral TKA, inflammatory arthritis, active infection, major neurologic disease affecting gait, recent lower-limb fracture, severe cognitive impairment, or inability to complete functional testing. Eighty eligible patients completing the required assessments formed the final cohort. Sample size. Assuming delayed or suboptimal postoperative recovery in approximately 25% of patients [1,2], a 95% confidence level and 10% absolute precision produced a minimum estimate of about 72 participants. After allowing approximately 10% for incomplete follow-up, the target sample was set at 80. Clinical assessment. Baseline variables included age, sex, body mass index (BMI), diabetes mellitus, hypertension, hemoglobin concentration, and varus deformity. Obesity was defined as BMI ≥30 kg/m². Functional status was measured with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), a validated osteoarthritis-specific instrument [10]; higher function-domain scores indicated greater disability. Objective mobility was assessed with the Timed Up and Go (TUG) test [11]. Preoperative WOMAC function and TUG values were recorded before surgery. Perioperative and follow-up assessment. Operative duration, postoperative day 1 numerical rating scale (NRS) pain score, hemoglobin decline, independent ambulation by postoperative day 3, hospital stay, and early complications were prospectively documented. WOMAC function and TUG were reassessed at 6 and 12 weeks. Active knee flexion and independent walking without an assistive device were recorded at 12 weeks. Delayed functional recovery was defined using a study-specific composite endpoint: achievement of fewer than two of three 12-week milestones—WOMAC function score <40, TUG ≤12 seconds, and independent walking without an assistive device. Patients achieving at least two milestones were categorized as having timely recovery. Statistical analysis. Continuous variables were summarized as mean ± standard deviation and categorical variables as frequency and percentage. Independent-samples t tests and chi-square or Fisher exact tests were used for between-group comparisons as appropriate. Variables associated with delayed recovery on univariable analysis and clinically relevant covariates were considered for multivariable binary logistic regression. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were reported. Discrimination was assessed by the area under the receiver operating characteristic curve; calibration was assessed using the Hosmer-Lemeshow test. A two-sided p value <0.05 indicated statistical significance. Ethical considerations. Necessary Permissions were obtained before starting the study. Written informed consent was obtained before enrolment, and the study was conducted in accordance with the Declaration of Helsinki.

A total of 80 patients with advanced knee osteoarthritis who underwent primary total knee arthroplasty (TKA) were included in the final analysis. The mean age of the study population was 66.8 ± 7.2 years, and 50 (62.5%) patients were female. The mean body mass index (BMI) was 29.1 ± 4.1 kg/m²; 31 (38.8%) patients were obese (BMI ≥30 kg/m²). Diabetes mellitus was present in 28 (35.0%) patients and hypertension in 42 (52.5%). The mean preoperative WOMAC functional score was 64.1 ± 11.5, while the mean preoperative TUG duration was 14.8 ± 3.6 seconds.

At the 12-week postoperative assessment, 24 patients (30.0%) demonstrated delayed functional recovery, whereas 56 (70.0%) achieved timely functional recovery. Patients with delayed recovery were significantly older than those with timely recovery (70.4 ± 6.1 vs. 65.3 ± 7.1 years; p = 0.003) and had a higher mean BMI (31.1 ± 4.0 vs. 28.3 ± 3.8 kg/m²; p = 0.004). Obesity and diabetes mellitus were also significantly more frequent among patients experiencing delayed recovery. Baseline characteristics according to functional recovery status are presented in Table 1.

Table 1. Baseline characteristics according to functional recovery status after total knee arthroplasty (n = 80)

Data are presented as n (%) unless otherwise specified. Higher WOMAC scores indicate greater functional limitation. BMI = body mass index; TUG = Timed Up and Go; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.

Significant differences were also observed in early postoperative variables. The mean pain intensity on the first postoperative day was higher among patients who subsequently experienced delayed recovery (6.3 ± 1.4 vs. 5.4 ± 1.3 on the numerical rating scale; p = 0.006). These patients required a longer hospital stay and were less likely to achieve independent ambulation by the third postoperative day. Postoperative complications were relatively uncommon, although their frequency was numerically greater among patients with delayed recovery (16.7% vs. 7.1%). The perioperative findings are summarized in Table 2.

Table 2. Perioperative characteristics according to functional recovery status

Data are presented as mean ± SD or n (%). NRS = numerical rating scale.

Functional improvement was observed during follow-up; however, the magnitude and speed of improvement differed substantially between the two groups. At 6 weeks, the mean WOMAC functional score remained significantly higher among patients with delayed recovery (52.8 ± 10.7) compared with those with timely recovery (35.6 ± 10.2; p <0.001). The difference persisted at 12 weeks, with corresponding scores of 44.9 ± 10.6 and 20.8 ± 8.7, respectively (p <0.001). Similarly, TUG performance improved more slowly in the delayed-recovery group. At 12 weeks, the mean TUG duration was 14.6 ± 3.1 seconds in the delayed-recovery group compared with 9.5 ± 2.2 seconds among patients with timely recovery (p <0.001).

The mean active knee flexion at 12 weeks was lower among patients with delayed recovery (102.6 ± 9.8° vs. 112.8 ± 8.6°; p <0.001). Furthermore, only 7 (29.2%) patients in the delayed-recovery group were able to walk independently without an assistive device at 12 weeks, compared with 48 (85.7%) patients in the timely-recovery group (p <0.001). Functional outcomes during follow-up are presented in Table 3.

Table 3. Functional outcomes during postoperative follow-up

Data are presented as mean ± SD or n (%). TUG = Timed Up and Go; WOMAC = Western Ontario and McMaster Universities Osteoarthritis Index.

On univariable analysis, age ≥70 years, obesity, diabetes mellitus, a higher preoperative WOMAC functional score, preoperative TUG >15 seconds, lower preoperative hemoglobin, and a postoperative day 1 pain score ≥6 were associated with delayed functional recovery. Clinically relevant variables were subsequently entered into the multivariable logistic regression model.

After adjustment for potential confounding factors, preoperative TUG >15 seconds emerged as the strongest independent predictor of delayed functional recovery (aOR 4.06; 95% CI 1.40-11.79; p = 0.010). Obesity was associated with approximately threefold higher odds of delayed recovery (aOR 3.18; 95% CI 1.12-9.02; p = 0.030). Age ≥70 years (aOR 2.91; 95% CI 1.03-8.19; p = 0.044), diabetes mellitus (aOR 2.87; 95% CI 1.01-8.17; p = 0.048), and postoperative day 1 NRS pain score ≥6 (aOR 3.51; 95% CI 1.20-10.29; p = 0.022) also remained independently associated with delayed recovery (Table 4).

Table 4. Multivariable logistic regression analysis of predictors of delayed functional recovery

OR = odds ratio; CI = confidence interval; BMI = body mass index; TUG = Timed Up and Go; NRS = numerical rating scale.

The final multivariable model demonstrated good discrimination for delayed functional recovery, with an area under the receiver operating characteristic curve of 0.84. The Nagelkerke R² value was 0.46, indicating that the included clinical variables explained approximately 46% of the variability in delayed recovery. The Hosmer-Lemeshow goodness-of-fit test was non-significant (p = 0.71), suggesting adequate model calibration. Overall, advanced age, obesity, diabetes, impaired preoperative mobility, and greater early postoperative pain were important predictors of delayed functional recovery following TKA.

The present prospective study found that 30.0% of patients experienced delayed functional recovery 12 weeks after primary TKA. Delayed recovery was characterized by poorer WOMAC function, slower TUG performance, reduced knee flexion, and greater dependence on walking aids. The principal independent predictors were preoperative TUG >15 seconds, obesity, age ≥70 years, diabetes mellitus, and greater pain on the first postoperative day. The multivariable model showed good discrimination, indicating that a combination of simple preoperative and early postoperative variables can identify patients at increased risk during rehabilitation. Preoperative mobility emerged as the strongest predictor. Patients with TUG >15 seconds had more than fourfold higher adjusted odds of delayed recovery. This agrees with earlier work showing that baseline functional limitation influences postoperative function [3,4] and that age and comorbidity affect acute physical recovery after TKA [5]. Ferreira et al. found that preoperative performance measures, including TUG, contributed to prediction of later functional outcomes [7]. More recently, Nakajima et al. reported a significant association between preoperative TUG time and 1-year Oxford Knee Score, with a proposed threshold of approximately 13 seconds for favorable outcome [13]. These observations support routine performance-based assessment before surgery. Obesity and diabetes were also independently associated with delayed recovery. Higher BMI has been incorporated into validated prediction models for poorer outcomes after knee replacement [6]. In the present cohort, obesity increased the adjusted odds of delayed recovery approximately threefold, while diabetes produced a similar increase. Lovie et al. reported that combined obesity and diabetes were associated with diminished knee-specific functional improvement after primary knee replacement [12]. Excess body mass increases mechanical demand during rehabilitation, whereas diabetes is frequently accompanied by metabolic and vascular factors that can hinder physical recovery. Long-term multicenter evidence has likewise identified BMI and comorbidity burden among predictors of functional outcome after TKA [14]. Early postoperative pain was another clinically relevant predictor. A postoperative day 1 NRS score ≥6 independently increased the odds of delayed recovery. Lo et al. showed that greater early pain after TKA was associated with poorer subsequent function and lower satisfaction [8]. A systematic review also identified moderate-to-severe acute postoperative pain as an important risk factor for persistent postsurgical pain [9]. Severe pain can restrict weight-bearing and participation in physiotherapy. In our cohort, delayed-recovery patients were less likely to ambulate independently by postoperative day 3 and had longer hospitalization, reinforcing the relationship between early postoperative barriers and slower functional progression. Age ≥70 years remained independently associated with delayed recovery after adjustment. This is consistent with studies linking age to early performance and longer-term function [5,7,14], although age associations vary according to whether the endpoint is pain, absolute function, or magnitude of improvement. The findings favor multidimensional risk assessment rather than age alone. Preoperative optimization, aggressive multimodal pain control, and targeted rehabilitation for high-risk patients represent practical strategies, while the prediction model requires external validation in larger cohorts. Limitations This study has several limitations. It was conducted at a single center with a modest sample size, which restricts external generalizability and precision of regression estimates. Functional recovery was assessed only through 12 weeks, so longer-term trajectories were not captured. The operational definition of delayed recovery combined performance and patient-reported measures and requires external validation. Residual confounding from psychosocial status, muscle strength, rehabilitation adherence, and socioeconomic factors also remains possible.

Delayed functional recovery occurred in 30.0% of patients during the first 12 weeks after primary total knee arthroplasty for advanced osteoarthritis. Preoperative TUG >15 seconds was the strongest independent predictor, while obesity, age ≥70 years, diabetes mellitus, and greater postoperative pain independently increased the likelihood of delayed recovery. Patients with delayed recovery demonstrated poorer WOMAC scores, slower mobility, reduced knee flexion, and greater dependence on walking aids. These findings support preoperative functional assessment and early identification of modifiable risk factors. Optimizing metabolic status, pain control, and rehabilitation intensity for high-risk patients could facilitate earlier functional gains. Larger multicenter studies with longer follow-up are required to validate the proposed prediction model.